中文摘要：

根据灰色理论，以轨道质量指数检测数据为原始时间序列，通过累加弱化序列的随机性，挖掘轨道系统内在的规律，研究建立基于灰色GM（1，1）非等时距模型的轨道质量预测方法。为提高模型预测精度，优化模型中的初值和背景值，并基于残差分析引入周期性函数，对模型进行修正。用此模型对轨道质量指数TQI数据进行分析预测，并对模型精度进行检验。结果表明模型能较好地反映轨道质量恶化发展的随机波动特征，拟合、预测精度高，为了解和掌握轨道质量状态的发展规律提供了新的方法。

英文摘要：

On the basis of grey theory, and with the inspection data of the track quality index as the original time series, by accumulating and weakening the randomness of sequences, and exploring the inherent law of the track system, a prediction method is established for track quality based on grey GM （1, 1） nonequal time interval correction model. To improve the prediction accuracy, the starting value and the background value in the model are optimized, and the periodic function is introduced based on residual analysis to correct the model. With this model, TQI （Track Quality Index） data are analyzed and predicted. Meanwhile the precision of the model has been verified. The results indicate that the model can well reflect the random fluctuation characteristics of track quality degradation development. With high fitting precision and high prediction accuracy, the model provides a new method for understanding and mastering the development law of the track quality status.